Hello-MCP 🚀

@Cookie-HOO

About Hello-MCP 🚀

A Simple MCP Demo With Client & Server

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "hello-mcp": {
      "command": "uv",
      "args": [
        "venv",
        "#",
        "Create",
        "virtual",
        "environment"
      ]
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is Hello-MCP 🚀?

Hello-MCP 🚀 is a simple implementation of an MCP (Model Control Protocol) client and server. It helps developers understand what the MCP protocol is and how it works by providing hands-on experience in building and using both client and server components.

How to use Hello-MCP 🚀?

Clone the repository, install uv (e.g., via curl -LsSf https://astral.sh/uv/install.sh | sh on macOS), create a virtual environment (uv venv), install dependencies (uv sync), and activate it. Configure your DeepSeek API key in config.yaml (copy from config.example.yaml). Run the client in stdio mode with python -m hello_mcp.client2stdio --server-path ./hello_mcp/server.py or in SSE mode by first starting the server (python -m hello_mcp.server --transport sse) then the client (python -m hello_mcp.client2sse --server-url http://127.0.0.1:8000). Debug the server using uv run mcp dev hello_mcp/server.py.

Key features of Hello-MCP 🚀

Supports registering MCP servers via SSE and stdio transports
Interactive command-line chat interface
Integrates with DeepSeek API (configurable API key)
Streamed response output
Simple server built on FastMCP with a path utility tool
Health check endpoint on the server

Use cases of Hello-MCP 🚀

Learning the MCP protocol by building your own client and server
Gaining hands-on experience with MCP tool registration and invocation
Debugging MCP servers using the built-in MCP Inspector
Experimenting with MCP in daily AI workflows (e.g., with Claude Desktop or Cursor)
Extending the server with custom tools (by adding functions with the @tool decorator)

FAQ from Hello-MCP 🚀

What problems does MCP solve?

MCP standardizes AI model interaction protocols, provides unified tool calling specifications, simplifies complex AI system integration, and enables modular, reusable model capabilities.

How do I set up the DeepSeek API key?

Copy config.example.yaml to config.yaml and fill in your DeepSeek API key (obtained from https://platform.deepseek.com/api_keys). Ensure no extra spaces or quotes around the key.

What if the port is already in use when starting the server?

Use lsof -i :8000 to find the process using the port, terminate it, or change the server port and update the client registration URL accordingly.

How can I extend Hello-MCP with new tools?

Add a new function in server.py, decorate it with @tool, define the input parameter schema, and test it by calling the client.

What are the runtime requirements?

Python 3, uv for project management, and a DeepSeek API key (currently the only supported AI provider). No additional MCP hosts are required; the server runs as a standalone Python process.

Frequently asked questions

What problems does MCP solve?

MCP standardizes AI model interaction protocols, provides unified tool calling specifications, simplifies complex AI system integration, and enables modular, reusable model capabilities.

How do I set up the DeepSeek API key?

Copy `config.example.yaml` to `config.yaml` and fill in your DeepSeek API key (obtained from https://platform.deepseek.com/api_keys). Ensure no extra spaces or quotes around the key.

What if the port is already in use when starting the server?

Use `lsof -i :8000` to find the process using the port, terminate it, or change the server port and update the client registration URL accordingly.

How can I extend Hello-MCP with new tools?

Add a new function in `server.py`, decorate it with `@tool`, define the input parameter schema, and test it by calling the client.

What are the runtime requirements?

Python 3, `uv` for project management, and a DeepSeek API key (currently the only supported AI provider). No additional MCP hosts are required; the server runs as a standalone Python process.